Why has Moloco's Retail Media solution experienced a 20% decline in return on ad spend for e-commerce clients since the latest algorithm update?

Introduction

Moloco's Retail Media solution has experienced a 20% decline in return on ad spend (ROAS) for e-commerce clients since the latest algorithm update. This significant drop in performance is concerning and requires a thorough investigation to identify the root cause and implement effective solutions. I'll approach this issue systematically, examining both internal and external factors that could be contributing to the decline.

Step 1

Clarifying Questions (3 minutes)

• Looking at the timing, I'm thinking the algorithm update might be directly responsible. Could you provide more details on the specific changes made in the latest update?

Why it matters: Understanding the changes helps pinpoint potential issues in the algorithm itself. Expected answer: Details on algorithmic changes, such as bid optimization or targeting adjustments. Impact on approach: If significant changes were made, we'd focus more on the algorithm; if minor, we'd explore other factors more heavily.

• Considering the scale of the decline, I'm wondering about the affected client base. Has this 20% decline been consistent across all e-commerce clients or are there variations?

Why it matters: Identifying patterns in affected clients could reveal specific vulnerabilities or issues. Expected answer: Information on whether the decline is uniform or varies among different client segments. Impact on approach: Variations would lead us to investigate client-specific factors more closely.

• Given the importance of seasonality in e-commerce, I'm curious about the timeframe of this decline. Over what period has this 20% decrease been observed?

Why it matters: Seasonal fluctuations could be mistaken for algorithm-related issues. Expected answer: The specific timeframe of the observed decline, ideally compared to the same period in previous years. Impact on approach: A short-term decline might indicate a temporary issue, while a longer-term trend would suggest more fundamental problems.

• Considering potential data discrepancies, has there been any change in how ROAS is calculated or measured since the algorithm update?

Why it matters: Ensures we're comparing apples to apples and not misinterpreting data due to measurement changes. Expected answer: Confirmation of consistent ROAS calculation methods or details of any changes. Impact on approach: If calculation methods changed, we'd need to re-evaluate the perceived decline before proceeding with other analyses.